Seattle's $20,000,000 Loss on its Private Equity Investment in Epsilon

Dear Mayor McGinn, Council Budget Chair, Tim Burgess, & Council Members

Subject: What Can Seattle Learn From the Epsilon II Calamity?

I’m writing in regard to the May 5, 2013 Seattle Times article entitled “Seattle City Pension Investment Mired in Legal Limbo,” which describes the regrettable misadventure of Seattle’s $20 million investment in the private equity fund, Epsilon II, which is now worthless.

A little more than a year ago I sent you a letter outlining the conflicts of interest that make the City's pension system vulnerable to the sort of investment losses described in the Times article, and I explained why Seattle’s continued practice of betting on investment managers who insist they can “beat the market” has a very low probability of success.

I was gratified to read Councilmember Burgess’s admission that “our private-equity investments have been very problematic, and Epsilon is clearly one of those.  We need to get a better handle on this part of our portfolio and better understand the risks as well as the rewards.”

This analytical commitment marks a significant improvement over the rather blasé attitudes expressed by some City officials, elected and unelected alike, regarding the City Pension System's investment practices.  The calamity of Epsilon II is, in fact, a symptom of a more general problem, which is the City’s continued reliance on financial consultants who, as a matter of temperament and self-interest, offer advice that is inconsistent with some of the most well established findings of financial economics.

In the Spring 2013 issue of the Journal of Economic Perspectives, there’s an article entitled “Asset Management Fees and The Growth of Finance,” in which Burton Makiel details the underperformance of actively managed investment funds compared to index funds. Moreover, the largest underperformance gap is in small capitalization stocks, where Seattle's financial advisors claim they can exploit market inefficiencies to achieve superior returns.  In light of these findings, which have been confirmed by many other studies, Makiel concludes,“Perhaps the greatest inefficiency in the stock market is in ‘the market’ for investment advice.”

Seattle's pension system has been a willing buyer in “the market for investment advice” and has paid dearly for it.  Between 1988-2011, Seattle's investments have underperformed a risk-equivalent index portfolio by 0.9%/year, not including fees.  This performance gap, which does not depend on hindsight, cost Seattle's pension system more than $500,000,000 (2011 dollars).  

According to the Times article, Seattle is considering the possibility of turning its investment decision making over to the Washington State Investment Board because the State’s investment performance has been superior to the City’s results.  But this is the wrong standard of comparison.  The performance of the Washington State Investment Board has been significantly worse than the performance of a risk-equivalent combination of index funds.  The last figure in the attachment displays the performance of 24 state pension funds versus 7 index portfolios for the 10-year period, 7/1/2000 to 6/30/2010. At every risk/return combination, the index portfolios outperformed the state pension funds by at least 2%/year. Washington State underperformed the risk-equivalent index portfolio by more than 2%/year.

The problem facing Seattle is that very few financial advisors would recommend an index portfolio because: 1) there’s more money to be made in active asset management than in the construction of an index portfolio; and 2) virtually all financial advisors believe they can outperform their competitors, but, unless they all live in Lake Wobegone, they can’t all be above average.  

It’s time for Seattle to consider a new approach to investing, one that draws on the mountain of academic research showing that active asset management is a losing strategy.  Continuation of the current approach will only deliver superior returns if Seattle's pension board can pick those few investment managers who can “beat the market,” a premise that implies a good deal of hubris, especially in light of the Epsilon II catastrophe.

Sincerely Yours, Greg Hill

Biting the Hand that Fed Me: UW Political Scientists on Deficit Reduction

Dear Professors Mark A. Smith and Rebecca Thorpe

I read with interest your brief essay, “Perspectives on National Government Spending,” in the Political Science Department’s Spring 2013 Newsletter.  Your point about the contradictory views of the public regarding “spending in general” vs. “spending on particular programs” is, indeed, helpful in understanding why it’s difficult to make significant cuts in government spending.

But several other claims you advance regarding the prospects for deficit reduction seem misleading if not mistaken. 

1.  To begin with, legislation adopted since the beginning of fiscal year 2011 reduced projected deficits by $2.4 trillion between 2013-2022, with 75% of this reduction attributable to spending reductions and interest savings.  These measures, which don’t include the recent sequester cuts, are forecast to reduce Federal debt as a percentage of GDP in 2022 from 93% to 83%.  In short, significant reductions in projected deficits have already been achieved.

2.  The proposition that long-term deficits “are driven mainly by entitlement spending” is correct, but misleading.  Social Security is not in bad shape; it can pay full benefits until 2033 with no changes to current law.  And Social Security’s long-term “problem” can be solved by modest changes phased in over time.  (See http://www.cbpp.org/cms/?fa=view&id=3261) Leaving Social Security aside, then, the remaining large entitlements, Medicare and Medicaid, are really symptoms of problems with the American health system.  Our per capita health care spending is, on average, twice that of other advanced economies, yet our health outcomes are worse on many dimensions.  The challenge is not to “reduce entitlements,” e.g., by increasing the Medicare eligibility age from 65 to 67, but to fix our health care system. 

3.  It’s not true that “both parties have resisted offering any cuts to entitlements.”  President Obama’s ACA reduced Medicare costs by over $700 billion.  And Obama has agreed to adopt the chained CPI for Social Security and to reduce subsidies for high-income Medicare recipients.  In addition, the “Stimulus Package” and the ACA include many investments and cost-effective practices that will reduce health care costs.

4.  Although it may be true that “Republicans seem more committed than Democrats to deficit reduction,” Republicans, in fact, aren’t more committed to deficit reduction (my stress).  Consider the following Republican policy choices: 1) the Bush tax cuts; 2) the unfunded Bush prescription drug benefit; 3) the two wars not paid for; 4) Republican unwillingness to raise taxes even though revenues as a percentage of GDP are very low by historical standards; and 5) the Paul Ryan budgets that include large tax cuts and only achieve balance by assuming implausible supply side elasticities.

5.  Finally, you seem to accept the argument that reductions in government spending will reduce the government’s deficit.  The problem is that cutting spending reduces income while increasing expenditure for unemployment benefits, food stamps, etc.  The disappointing effects of the spending reductions in Europe and the U.K. on public sector deficits raise serious questions about the virtues of austerity.  You might find this post interesting: http://mainlymacro.blogspot.com/2013/03/why-politicians-ignore-economists-on.html

Sorry to drone on.  I do appreciate your willingness to write for a general audience and, as you can tell, I found your piece thought provoking.

Best regards, Greg Hill (Ph.D. in Political Science, UW, 1982 (or 1981, I can’t remember)

Stephen Williamson Does *Not* Sort Out Keynesian Economics

Stephen Williamson took a break from haranguing Paul Krugman to post a working paper about Keynesian economics.  A newer version of the paper, entitled “Sorting Out Keynesian Economics: A New Monetarist Approach,” can be found on Professor Williamson’s academic website here.  (All citations below are from the newer working paper.)

Williamson is puzzled by the persistence of Keynesian economics given the “revolution in thought” that took place after 1970. “However, to be fair to the Keynesians,” Williamson allows, “we should at least try to understand what they are up to, and evaluate the work,” an evaluation which culminates in Williamson's judgment that “there is nothing about [my] paper that should make us any more comfortable with Keynesian analysis.  All of the basic questions remain unanswered." 

Williamson claims his model is closest to the model developed by Roger Farmer in his 2012 Economic Journal article, "Confidence, Crashes, and Animal Spirits."  

“We can imagine a world – Farmer’s Keynesian world - where a matched worker and producer in our model have difficulty splitting the surplus.  Then, there exists a continuum of equilibria, indexed by wages and labor market tightness.  In general, an equilibrium with a high (low) wage is associated with low (high) labor market tightness . . . In this static model, the equilibrium can be suboptimal, in that labor market tightness is too high or too low. Indeed, the unemployment rate could be too high or too low.”

Later, Williamson modifies his simple search model to include a dynamic framework, consumers, and money.  Throughout these refinements, Williamson’s basic conclusion holds:

“Inefficiencies arise in Keynesian models because private sector economic agents somehow do not get the terms of exchange right.  In our model, what goes wrong is that there are no incentives determining how producers, workers, and consumers split up the gains from trade.”

Although this putative market failure seems to invite corrective policy, a successful policy, according to Williamson, requires civil servants who can see things that aren’t visible to market participants.  

“As in typical Keynesian policy analysis, those agents who are assigned the job of working out optimal monetary and fiscal policy somehow see through these inefficiencies and are able to design policies that correct the problems.  In this sense, our analysis is no more, and no less, sensible than what the average Keynesian does.” 

 Quite a devastating assessment it would seem.  Yet, in spite of the paper’s self-assured tone, there’s much less here than meets the eye.  To begin with, Williamson’s characterization of Farmer’s model as one “where a matched worker and producer . . . have difficulty splitting the surplus” actually bears very little resemblance to Farmer’s Old Keynesian model.  Here’s Farmer’s own characterization,

“In this article, I propose a new approach.  Instead of searching for a fundamental explanation to close an indeterminate model of the labour market, I close the model with the assumption that firms produce as many goods as are demanded. Demand, in turn, depends on beliefs of market participants about the future value of assets.  By embedding the indeterminate labour search market into an asset pricing model, I show that the unemployment rate can be explained as a steady-state equilibrium where the indeterminacy of equilibrium is resolved by assuming that the beliefs of market participants are self-fulfilling.”

In other words, Farmer assumes: 1) firms produce goods to meet demand; 2) demand, in turn, depends on beliefs about the future value of assets; 3) these beliefs are self-fulfilling, which means; 4) the unemployment rate is determined by self-fulfilling beliefs about future asset values.  Thus, unemployment in Farmer’s model, far from being the consequence of “not getting the terms of trade right,” as Williamson suggests, is, in fact, due to a lack of aggregate demand that’s rooted in self-fulfilling pessimism about future asset values.  In short, Williamson has stripped away the most Keynesian element in Farmer’s model, viz. the role of beliefs, or “animal spirits,” in determining the level of employment via their effect on asset values, which, in turn, influence demand.

The bearing of Williamson’s modeling exercise on “Keynesian economics” is also put in question by the particular search model he’s chosen to work with.  In Williamson’s model, producers, workers, and consumers must find one another and then come to agreement regarding the “terms of exchange.”  This agreement is the subject of bargaining, which, in the absence of certain constraints, won’t produce an efficient outcome.  

By contrast, Farmer builds on Peter Diamond’s search model, where an increase in the availability of potential trading partners raises the profitability of engaging in trade, thereby generating externalities and multiple equilibria.  In fact, Farmer sees “no reason to treat the search model differently from any other competitive model with externalities,” adding that he views “the addition of the bargaining equation as arbitrary.”  So, the externalities that are the centerpiece of the search model Farmer employs have also disappeared in Williamson’s version of Farmer’s model.

We are now in a better position to evaluate Williamson’s conclusion:

“There is nothing about this paper [i.e., Williamson’s own paper] that should make us any more comfortable with Keynesian analysis.  All of the basic questions remain unanswered.  Why are private sector agents so bad at determining the terms of exchange, while public sector agents are so good at it?  How could such seemingly small difficulties in setting prices and wages lead to such large aggregate problems with the allocation of resources?”

Not to put too fine a point on it, but the reason there’s nothing in Williamson’s paper “that should make us any more comfortable with Keynesian analysis” is because the paper has very little to do with Farmer’s Old Keynesian model (and even less to do with Keynes’ own theory).

Brad Delong, Thomas Nagel, and The Relationship between Mind and World

Brad Delong has a provocative post here on Thomas Nagel’s new book, Mind and Cosmos: Why the Materialist Neo-Darwinian Conception of Nature Is Almost Certainly False.  Delong dismisses Nagel’s central argument, but that’s because Brad doesn’t really grasp Nagel’s point of view, or even the point of philosophy (at least in this post).  Delong attributes the following line of reasoning to Nagel, 

Suppose we think we are going south-southwest and see the sun rising before us. [According to Nagel] we don't think: "the heuristics of reasoning that have evolved because they tend to boost reproductive fitness make it likely that I am not in fact going south southwest". We think, instead: "I know the sun rises in the east! I must be going roughly east! I deduce this by my reason, and my reason is a mechanism that can see that the algorithm it follows is truth-preserving! My mind is in immediate contact with the rational order of the world! I don't just think I am going east! I know I am going east!

Delong evidently thinks Nagel is trying to answer the question, “why do we think we’re right when we hold a belief with great confidence?” when, in fact, Nagel is asking whether our minds are something more than mechanisms designed to “boost reproductive fitness.”

Delong proceeds to show, with understandable ease, that reasoning creatures like ourselves can be wrong even when we’re absolutely convinced we’re right.  So, Brad concludes

Any theory that provided such an account of reason becoming an instrument of transcendence and offering guarantees of grasping objective reality would be hopelessly, terribly, laughably wrong.

Now, if you think Nagel is arguing against the very possibility of mistaken belief, then I suppose you might conclude, as Brad does, that Nagel is “hopelessly, terribly, laughably wrong,” (though, if you’d read a lot of Nagel’s work, you’d probably be more charitable).

But when Delong applies the coup de grace, saying

I cannot help but think that only a philosophy professor would believe that our reason gives us direct access to reality.  Physicists who encounter quantum mechanics think very differently. . .

. . .  Brad inadvertently illuminates the question on which Nagel should have been focused, as well as Delong’s own unarticulated and debatable understanding of the relationship between philosophy and science.  The point of philosophy is, in part, to clarify what it means to have “access to reality,” a task that requires reflection on several issues central to Western philosophy from its classical beginnings in ancient Greece.

In Delong’s view, at least as it’s expressed in his post, physics is the ultimate arbiter on questions concerning the nature of reality.  And this understanding has the implication, I think it’s fair to say, that it’s up to science to tell us what counts as having “access to reality.”  W.V.O. Quine made a strong case for a conception that resembles my reading of Delong, but Quine was unsuccessful in trying to assimilate philosophy to science, and to destroy, along the way, the autonomy of philosophical (and ordinary) reason.

For Quine, the world stands to thought only as cause to effect, and, as Richard Rorty puts it in commending this view, “we understand all there is to know about the relation of beliefs to the world when we understand their causal relations to the world.”  The trouble is that this understanding rules out the possibility that our thought stands in relation to reality in such a manner that features of the world can provide reasons for a particular judgment or belief.

Here’s Wittgenstein’s way of putting it, “When we say, and mean, that such-and-such is the case, we – and our meaning – do not stop anywhere short of the fact; but we mean: this-is-so.”  Wittgenstein is not, of course, claiming that all our beliefs are true, but rather that there’s no conceptual gap between the sort of thing that can be thought (or said) and the sort of thing that can be the case in the world.  That things are thus-and-so is both the content of experience and (if we’re not mistaken) an aspect of the world.

From this vantage point, the “materialist neo-Darwinian conception of nature” isn’t mistaken because it has failed to produce a theory that explains the power of mind to comprehend the world as it is.  It’s mistaken because it regards the world only as the cause of beliefs and not as an object of beliefs, which can be assessed according to the standards appropriate to the particular beliefs in question.

How to Solve Our Massive Macroeconomic Coordination Failure

Robert J. Barro asked me here to identify the market failure that would justify the American Recovery and Reinvestment Act (aka “the stimulus”).  My answer, in brief, is that the market failure which permits sustained high unemployment is the lack of a truly comprehensive market in which workers with excess leisure, and firms with excess capacity, could make conditional commitments of the following kind, “if you’ll hire me at the going wage, I’ll promise to purchase goods from firms A, B, and C, whose new hires, in turn, will promise to buy goods from you.”

In the absence of such a market, a firm’s decision to expand employment will give rise to a positive externality insofar as the new hires buy the output of other firms.  (I assume the new hires were unemployed, and that prices exceed marginal costs, so aggregate demand rises, and firms profit from additional sales.)  Since firms don’t take account of this positive externality in their hiring decisions, total employment will be less than optimal.

At first glance, persistent unemployment looks like a coordination problem.  Imagine a two-firm economy in which the wages paid out by firm X are spent on the output of firm Y, and vice versa. If X believes Y will be hiring additional labor, then X will expand its output, and if Y believes the same about X, Y will expand, and the result will be a high-output equilibrium. If both X and Y expect the other to reduce output, then you’ve got a low-output equilibrium.  The trick, in this case, is to get X and Y to coordinate on the high-output equilibrium.

Laurence Kotlikoff has suggested that coordinating around high employment could be accomplished if President Obama urged the largest 1,000 firms to hire additional employees, then urged the next largest 1,000 firms to expand employment, and so on.   

I don't believe this approach will work for the following reason: in game-theoretic jargon, this is really a Prisoner's Dilemma Game in which each firm's dominant strategy is to wait for other firms to start hiring.  It's too risky to expand employment unless you can be assured that other firms will do likewise.  By waiting, a firm avoids the risk of expanding employment without an increase in demand.  And the losses in this scenario would certainly be greater than the profits that would be foregone "by coming late to the party," i.e., hiring after the recovery has gained momentum.  So President Obama urging thousands of firms to expand employment can only be successful if a) Obama inspires an awful lot of confidence and/or b) firms have some altruistic component in their decision-making objectives.  These are both unreasonable expectations in my view.

Nevertheless, Professor Kotlikoff is onto something.  Our current unemployment problem is a coordination problem.  But you can't solve it without converting it into an Assurance Game in which each firm is assured that, if it hires, others will also hire.  In an earlier post, I suggested that large firms submit “offers” to a central auctioneer of the following kind: I’ll increase employment at my firm by X% if the total increase in employment promised by all other firms amounts to at least Y%.  This Walrasian process of tâtonnement would continue until the greatest increase in total employment commitments were achieved.  These hiring pledges, in turn, would be enforced by levying tax penalties on firms that failed to meet their commitments, and transferring revenue from these payments to firms that met their hiring commitments.

One drawback of this process is that firms would “low ball” their hiring commitments, understating the employment commitments they would really be willing to make.  To address this problem, the government could offer tax subsidies to those firms whose employment commitments exceeded the median employment commitment, measured by (say) the percentage increase over a firm’s actual employment in the previous year.   Alternatively, you could replace the median threshold with (say) the top quartile, etc.  Finally, the government could stipulate that it would only impose the tax penalties and transfer revenues if (say) 80% of the firms made hiring commitments.